Preti Antonio, Vellante Marcello, Petretto Donatella R
a Psychiatry Branch , Centro Medico Genneruxi , Cagliari , Italy.
b Section of Clinical Psychology, Department of Education, Psychology, and Philosophy , University of Cagliari , Cagliari , Italy.
Cogn Neuropsychiatry. 2017 May;22(3):233-253. doi: 10.1080/13546805.2017.1300091. Epub 2017 Mar 13.
The "Reading the Mind in the Eyes" Test (hereafter: Eyes Test) is considered an advanced task of the Theory of Mind aimed at assessing the performance of the participant in perspective-takingthat is, the ability to sense or understand other people's cognitive and emotional states. In this study, the item response theory analysis was applied to the adult version of the Eyes Test.
The Italian version of the Eyes Test was administered to 200 undergraduate students of both genders (males = 46%). Modified parallel analysis (MPA) was used to test unidimensionality. Marginal maximum likelihood estimation was used to fit the 1-, 2-, and 3-parameter logistic (PL) model to the data. Differential Item Functioning (DIF) due to gender was explored with five independent methods.
MPA provided evidence in favour of unidimensionality. The Rasch model (1-PL) was superior to the other two models in explaining participants' responses to the Eyes Test. There was no robust evidence of gender-related DIF in the Eyes Test, although some differences may exist for some items as a reflection of real differences by group.
The study results support a one-factor model of the Eyes Test. Performance on the Eyes Test is defined by the participant's ability in perspective-taking. Researchers should cease using arbitrarily selected subscores in assessing the performance of participants to the Eyes Test. Lack of gender-related DIF favours the use of the Eyes Test in the investigation of gender differences concerning empathy and social cognition.
“读心术眼测试”（以下简称：眼测试）被认为是心理理论的一项高级任务，旨在评估参与者在换位思考方面的表现，即感知或理解他人认知和情绪状态的能力。在本研究中，项目反应理论分析被应用于眼测试的成人版本。
对200名男女本科生（男性占46%）进行了意大利语版的眼测试。采用修正平行分析（MPA）来检验单维性。使用边际极大似然估计将单参数、双参数和三参数逻辑斯蒂（PL）模型拟合到数据中。采用五种独立方法探讨了性别导致的项目功能差异（DIF）。
MPA提供了支持单维性的证据。在解释参与者对眼测试的反应方面，拉施模型（单参数逻辑斯蒂模型）优于其他两个模型。尽管某些项目可能存在一些差异，反映了不同组之间的实际差异，但在眼测试中没有确凿证据表明存在与性别相关的DIF。
研究结果支持眼测试的单因素模型。眼测试的表现由参与者的换位思考能力决定。研究人员在评估参与者对眼测试的表现时应停止随意使用选定的子分数。缺乏与性别相关的DIF有利于在关于同理心和社会认知的性别差异研究中使用眼测试。
